Our measure of transit ridership in this study derives from the Census Transportation Planning Package (CTPP) compiled from the results of the 2000 Census. We are specifically interested in understanding the socioeconomic, transit service quality, and land use variables that influence the number of journey to work (commute) transit trips traveling from one zone (the origin zone) to another (the destination zone). The relevant dependent variables are available in Part 3 of the CTPP, which examines travel flows between pairs of TAZs for all reported commute trips. Travelers are assigned to one mode of transportation based on their responses to Census long form questions. Two transit modes are of interest for the Atlanta study: “bus or trolley bus” transit and “subway or elevated” rail transit, as these are the two modes in operation in Atlanta at the time of the 2000 Census.
The CTPP allows a transit rider to only choose one mode for their trip. For Atlanta, this mode is bus or rail. Many real-world transit trips involve the use of multiple modes, such as walk to bus, transfer to rail, transfer to bus and walk to rail, and transfer to bus. Many survey respondents undoubtedly make these and other multimodal trips, and it is unclear how they would categorize themselves. Would they refer to themselves as bus riders or rail riders? It is unclear. This is a limitation of the original CTPP data.
As we describe below, many individuals who identified themselves as bus riders in the CTPP data in reality transferred from bus to train and perhaps back again in making their trips to work. We therefore make adjustments, where appropriate, to their travel times as we discuss in the methodology section of the report to reflect instances of shorter travel times by combinations of buses and trains in comparison to bus-only trips. In terms of individuals who identified themselves as rail riders, and who may have accessed rail by automobile or bus, we also use the ARC On-Board survey to apportion rail riders across these potential access modes.71 For many rail riders, we calculate the access time to transit as a weighted proportion of the access times of the potential access modes, as we discuss later in the report.
Table 5 shows the destinations of transit riders using bus or train to get to work. The table shows that the types of destinations characteristic of the two groups of riders are very different and reflect the objectives of the bus route restructurings that accompanied the progressive introductions of rail transit segments. Before the route restructurings, long bus routes connected most neighborhoods in the Atlanta region to the Atlanta CBD. Rail lines substituted for much of the bus mileage running along trunk roads entering the CBD, and many of the shortened bus routes were terminated at rail stations, essentially forcing a transfer for those bus riders who wanted to reach the CBD. At the same time, however, different bus routes congregating at suburban rail stations made it possible for bus users to reach many suburban destinations.
Case Study Selection 37
Table 5. Destinations of Transit Riders Using Bus or Rail (2000)
Table 5 shows that in 2000 most bus riders (8,915 out of 12,371 riders surveyed) were in fact taking advantage of the suburban transfer capability, destined to suburban rather than CBD jobs. Many of the suburban jobs were located in job clusters, but a substantial number were scattered as well. Of the 8,915 bus riders traveling to suburban jobs, more than half of them (4,609) were traveling to what ARC defines as regional centers. About a quarter of the daily bus riders traveled to jobs in the CBD and the adjoining city center areas.
In contrast to bus riders, most rail riders (3,427 out of 6,121 riders surveyed) were headed to jobs in the CBD, and another 935 riders were destined to city center stations just north of the CBD. Still, there was a substantial number of rail riders headed to jobs in the rest of the metropolitan area (1,759 riders or 29 percent of rail riders surveyed), and most of these were headed to jobs in what ARC calls regional centers.
Figures 20 and 21 give greater meaning to these numbers. Figure 20 shows the destinations of bus riders going to work in the year 2000. The densest clusters of jobs accessed by bus riders were in the Buckhead area (in some cases close to rail stations), around Emory University and Decatur, around Georgia Tech, and along the south rail corridor toward the airport. Regional centers that had no or few jobs accessed by bus
riders typically had no bus service going to them (Gwinnett and Clayton Counties) or poor bus service (Cobb County).
Figure 20. Destinations of Atlanta Bus Riders Going to Work, 200072
Note: Zones colored using yellow-brown colors have transit service.
Figure 21 shows destinations of rail riders going to work in the year 2000. It shows that most rail riders were traveling to jobs immediately adjacent to rail stations, not only in the CBD and Midtown areas, but elsewhere. In some cases, such as stations in the CBD, stations serving Midtown, Lindbergh, and the Airport, many riders could walk to destinations. Other destinations of rail riders, such as Georgia Tech, a zone just north of Emory University, a freeway-oriented commercial area north of the city center adjacent to the split of I-75 and I-85, required riders to transfer to buses for the final leg of their trips. So, while most rail riders who were going to work outside of the CBD and city center were going to regional centers, it was only to a small subset of regional centers directly adjacent to rail stations or reached by a short bus ride. Rail riders tended to avoid traveling to regional centers located farther away, and there were a couple of regional centers located along the south rail line that they avoided, as well.
Case Study Selection 39
Figure 21. Destinations of Atlanta Rail Riders Going to Work, 200073
Note: Zones colored using yellow-brown colors have transit service.
Figure 22 shows the zones where work-bound bus riders lived in the year 2000. The great majority of riders originated in the southern and western districts of the metropolitan area, generally characterized by lower median household incomes, lower levels of vehicle accessibility, higher unemployment rates, and higher vacancy rates. Fewer bus riders came from more affluent zones. Some bus riders lived in zones along I-85 in the northeastern sector of the area as well as in central Cobb County, but these zones were also characterized by lower incomes. (See Figure 8, discussed earlier.)
Figure 22. Origins of Atlanta Bus Riders Going to Work, 200074
Note: Zones colored using yellow-brown colors have transit service.
Work-bound rail riders, on the other hand, lived in more dispersed locations in the year 2000, as shown in Figure 23. One reason that rail riders were more dispersed than bus riders is that about 52% of them accessed rail stations by auto, according to an onboard MARTA survey of riders.75 While conducted in 2002, the access patterns revealed by the survey were unlikely much different two years earlier. The census-based data in Figure 23 show that some work-bound rail riders live in a corridor along the south line, depressed by most economic measures, but others live in more affluent minority areas in the southeast and east of DeKalb County and others in moderate income areas of north-central Fulton County (see Figures 8, 11, and 12 shown earlier). In addition there is a scattering of riders living in more affluent areas throughout the region at great distances from the rail lines, a pattern indicating access by auto.
What can we make of these patterns? First, what emerges from the bus data is a group of riders from less affluent areas traveling to work in centers located largely in more affluent areas. Given that many of the destinations of these bus riders were adjacent to rail stations and that the transit route structure made it difficult to get to such areas except by transferring to MARTA trains, we suspect that many of the bus riders were in fact using a combination of bus and rail modes to complete their trips. One may infer that the bus and bus/rail service was a help economically to the population using it, and that the dispersed
Case Study Selection 41
nature of the bus system was a great help for this group in reaching the very dispersed distribution of jobs in Fulton and DeKalb Counties. Unfortunately, there remained many jobs without transit access outside of Fulton and DeKalb Counties.
Figure 23. Origins of Atlanta Rail Riders Going to Work, 200076
Note: Zones colored using yellow-brown colors have transit service.
The rail riders represent more of a mix of different types of riders. The figures indicate that some are similar in characteristics to bus riders, but there are clearly important differences as well. There appear to be significant numbers of more middle class rail riders with higher incomes and greater access to cars. A large part of this group may live in minority areas.
There are also some rail riders who appear to come from more affluent areas. In general, these riders appear willing to use trains to stations, from which they can walk to jobs directly; a minority of riders appear willing to transfer to buses to reach jobs located near stations.
To gain more insight about characteristics of transit riders and policies that might induce more of them to use transit, we now turn to specification and estimation of demand models.
43
IV. METHODOLOGY
The methodology in this study is to specify and estimate models that predict bus and rail transit work trips (the dependent variable) from one part of a metropolitan area (traffic analysis zone or TAZ) to another. In other words, we develop statistical equations that allow us to explain the influence of different types of variables on transit ridership. The approach is similar to that used by the authors in their statistical analysis of bus passenger demand in Broward County, Florida.77 In this case study, however, we employ two sets of models. For one set of models, the dependent variable consists of transit users who identified themselves as “bus or trolley bus” riders in the 2000 CTPP. For the other set of models, the dependent variable consists of transit users who identified themselves as
“subway or elevated” riders in the 2000 CTPP. We treat the former group as (self-identified) bus riders and the latter group as (self-identified) rail riders, although many riders in either category undoubtedly use multiple modes for their trips. The explanatory variables used in the models include socioeconomic variables from the 2000 CTPP, land use variables defined by the local MPO, and variables that measure transit service quality (broken into three components: in-vehicle, out-of-vehicle, and transfer time) obtained from the travel time skims of the regional travel demand model.